\begin{table}[ht!]
\centering
\footnotesize
\begin{threeparttable}\label{regressions_land}
\caption{Soil characteristics and farm size in 1852.} \label{regressions_land}
\begin{tabular}{lcccc}
\toprule
\toprule
 & (1) & (2) & (3) & (4)  \\
 \midrule
 \addlinespace
  \multicolumn{5}{l}{\textit{Panel A. Dep. Variable is Arable hectares per landowner}}  \\
  \addlinespace
\midrule
    \addlinespace
\ \ \ \ Ruggedness (100m)  & $-0.92$*** & & $-1.13$*** & $-2.10$***  \\
     & (0.32)         & & (0.37)       & (0.63)      \\   
  \addlinespace
\ \ \ \ Share sandy (0\textendash1) & & 7.37*  & 8.01*  & 9.62** \\
 & & (4.23) & (4.10) & (4.19) \\
   \addlinespace
\ \ \ \   Adjusted R-squared & 0.283 & 0.304 & 0.316 & 0.341 \\
    \addlinespace
    \midrule
 \addlinespace
  \multicolumn{5}{l}{\textit{Panel B. Dep. Variable is Cultivated hectares per landowner}} \\
  \addlinespace
\midrule 
    \addlinespace
    \ \ \ \ Ruggedness (100m)  & $-0.67$***& & $-0.80$*** & $-1.53$*** \\
 & (0.204)      & & (0.24)         & (0.43) \\
   \addlinespace
\ \ \ \ Share sandy (0\textendash1)   & & 4.51   & 4.96*** & 6.06** \\
                                               & & (2.80) & (2.71)   & (2.77) \\
        \addlinespace
\ \ \ \   Adjusted R-squared           & 0.252 & 0.268 & 0.284 & 0.311 \\
       \addlinespace
    \midrule
   Département FE & Y & Y & Y & Y  \\
Controls & . & . & . & Y  \\
N observations & 354 & 354 & 354 & 354  \\
N clusters         & 84    & 84  & 84 &  84  \\
\bottomrule
\end{tabular}
\begin{tablenotes}[flushleft]
\setlength\labelsep{0pt}
\item \footnotesize{\textit{Notes:} This table examines the relationship between soil characteristics and land concentration. The dependent variable is the average number of arable hectares per landowner of farming land in Panel~A and the average number of cultivated hectares per landowner of farming land in Panel~B. The sample is arrondissements in the 1852~agricultural census. Controls include an indicator variable for whether an arrondissement is on the coast, an indicator variable for whether it is mountainous, and an indicator variable for whether it is crossed by one of France's main rivers. Standard errors in parentheses are clustered by département. Estimates are calculated using Correia's \citeyear{reghdfe} \texttt{reghdfe} Stata command. *p$<$.10; **p$<$.05; ***p$<$.01.}
\end{tablenotes}
\end{threeparttable}
\end{table}